Research on Wireless Spectrum Sensing Technology Based on Machine Learning

  • Heng XiaoEmail author
  • Xianchun Zhou
  • Yue Tian
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11342)


In this paper, based on the spectrum sensing problem of the primary user signal in low SNR environment, an improved random forest spectrum sensing algorithm is proposed by using the advantage of strong forest classifier. The algorithm uses the mean and variance of the signal cycle spectrum with the largest extraction energy at each cycle frequency as the characteristic parameter to generate the samples in the sample set; On the basis of this, some samples in the presence of the primary user are selected as positive samples, and some samples in the absence of the primary user are used as negative samples to realize the construction of random forests; Finally, the trained random forests are used to classify the detected signals to achieve effective perception of the primary user signals. Thereby improving the perceptual performance of the main user signal in the case of low signal to noise ratio. Thereby improving the perceptual performance of the main user signal in the case of low signal to noise ratio. The experimental results show that the proposed algorithm has strong classification detection effect, can achieve spectrum sensing of the main user signal better under low SNR, and is generally suitable for solving the spectrum sensing problem of primary user signal in low SNR environment.


Spectrum sensing Primary user Random forest Decision tree 



This paper is funded by the following project funds: The Natural Science Foundation of Hainan Province (No. 617182, 618MS083), Sanya City Institute of Science and Technology Cooperation Project (No. 2015YD11, No. 2015YD57).


  1. 1.
    Xiao, H., Lv, S.: Efficient spectrum sensing mechanism with interference cancellation in cognitive radio network. J. Comput. Appl. 34(5), 1243–1246 (2014)Google Scholar
  2. 2.
    Sun, S., Kadoch, M., Gong, L., et al.: Integrating network function virtualization with SDR and SDN for 4G/5G networks. IEEE Netw. 29(3), 54–59 (2015)CrossRefGoogle Scholar
  3. 3.
    Pourgharehkhan, Z., Taherpour, A., Sala-Alvarez, J., et al.: Correlated multiple antennas spectrum sensing under calibration uncertainty. IEEE Trans. Wirel. Commun. 14(12), 6777–6791 (2015)CrossRefGoogle Scholar
  4. 4.
    Xiong, G., Kishore, S., Yener, A.: Spectrum sensing in cognitive radio networks: performance evaluation and optimization. Phys. Commun. 9, 171–183 (2013)CrossRefGoogle Scholar
  5. 5.
    Xiao, H., Lv, S.: Parallel channel competition mechanism in wireless network based on frequency domain. Comput. Eng. 41(8), 89–94 (2015)Google Scholar
  6. 6.
    Wang, X., Wang, J., et al.: Primary user signal recognition algorithm based on random forest in cognitive network. J. Northeast. Univ. (Nat. Sci.) 12, 1706–1709 (2015)zbMATHGoogle Scholar
  7. 7.
    Wang, X., Wang, J.-k., et al.: Spectrum sensing algorithm based on random forest in cognitive network. Chin. J. Sci. Instrum. 34(11), 2471–2477 (2013)Google Scholar
  8. 8.
    Xue, M., Li, Y.: An activity recognition algorithm on smart environment based on improved random Forest algorithm. Comput. Eng. 44(4), 1–8 (2018)MathSciNetGoogle Scholar
  9. 9.
    Wen, B., Dong, W., Xie, W., et al.: Parameter optimization method for random forest based on improved grid search algorithm. Comput. Eng. Appl. 54(10), 154–157 (2018)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  1. 1.Sanya UniversitySanyaChina
  2. 2.School of Information Science and Electronic EngineerHunan UniversityChangshaChina

Personalised recommendations